Whitesides Group: Writing a Paper
George M. Whitesides
Department of Chemistry and Chemical Biology, Harvard University, Cambridge, MA02138, USA
1.What is a scientific paper?
A paper is an organized description of hypotheses, data and conclusions, intendedto instruct the reader. Papers are a central part of research. If your researchdoes not generate papers, it might just as well not have been done.“Interesting and unpublished” is equivalent to “non-existent.”
Realize that your objective in research is to formulate and test hypotheses, todraw conclusions from these tests, and to teach these conclusions to others.Your objective is not to “collect data.”
A paper is not just an archival device for storing a completed researchprogram, it is also a structure for planning your research in progress. If youclearly understand the purpose and form of a paper, it can be immensely usefulto you in organizing and conducting your research. A good outline for the paperis also a good plan for the research program. You should write and rewritethese plans/outlines throughout the course of the research. At the beginning,you will have mostly plan; at the end, mostly outline. The continuous effort tounderstand, analyze, summarize, and reformulate hypotheses on paper will beimmensely more efficient for you than a process in which you collect data andonly start to organize them when their collection is“complete.”
2.1 The reason for outlines.
I emphasize the central place of an outline in writing papers, preparingseminars, and planning research. I especially believe that for you, and for me,it is most efficient to write papers from outlines. An outline is a writtenplan of the organization of a paper, including the data on which it rests. Youshould, in fact, think of an outline as a carefully organized and presented setof data, with attendant objectives, hypotheses and conclusions, rather than anoutline of text.
[按照提纲进行写作是最有效的方法][提纲是一篇论文的行文计划，应该包括论文所依靠的数据。提纲不仅仅是列出各段的内容，而是按照目的， 假说， 结论来精心组织数据。]
An outline itself contains little text. If you and I can agree on the detailsof the outline (that is, on the data and organization), the supporting text canbe assembled fairly easily. If we do not agree on the outline, any text isuseless. Much of the time in writing a paper goes into the text; most of thethought goes into the organization of the data and into the analysis. It can berelatively efficient to go through several (even many) cycles of an outlinebefore beginning to write text; writing many versions of the full text of apaper is slow.
All the writing that I do - papers, reports, proposals (and, of course, slidesfor seminars)- I do from outlines. I urge you to learn how to use them as well.
2.2 How should you construct an outline?
The classical approach is to start with a blank piece of paper, and write down,in any order, all important ideas that occur to you concerning the paper. Askyourself the obvious questions:“Why did I do this work?” “What does itmean?”“What hypothesis did I mean to test?”“What ones did I actuallytest?”“What were the results?”“Did the work yield a new method or compound?What?”“What measurements did I make?”“What compounds? How were theycharacterized?” Sketch possible equations, figures, and schemes. It isessential to try to get the major ideas written down. If you start the researchto test one hypothesis, and decide, when you see what you have, that the datareally seem to test some other hypothesis better, don't worry. Write them bothdown, and pick the best combinations of hypotheses, objectives and data. Oftenthe objectives of a paper when it is finished are different from those used tojustify starting the work. Much of good science is opportunistic andrevisionist.
When you have written down what you can, start with another piece of paper andtry to organize the jumble of the first one. Sort all of your ideas into threemajor heaps (A-C).
Why did I do the work? What were the central motivations and hypotheses?
B) Results and Discussion
What were the results? How were compounds made and characterized? What wasmeasured?
What does it all mean? What hypotheses were proved or disproved? What did Ilearn? Why does it make a difference?
Next, take each of these sections, and organize it on yet finer scale.Concentrate on organizing the data. Construct figures, tables, and schemes topresent the data as clearly and compactly as possible. This process can be slow- I may sketch a figure 5-10 times in different ways, trying to decide how itis most clear (and looks best aesthetically).
Finally, put everything—outline of sections, tables, sketches of figures,equations - in good order.
When you are satisfied that you have included all the data (or that you knowwhat additional data you intend to collect), and have a plausible organization,give the outline to me. Simply indicate where missing data will go, how youthink (hypothesize) they will look, and how you will interpret them if yourhypothesis is correct. I will take this outline, add my opinions, suggestchanges, and
return it to you. It usually takes 4-5 repeated attempts (oftenwith additional experiments) to agree on an outline. When we have agreed, thedata are usually in (or close to) final form (that is, the tables, figures,etc., in the outline will be the tables, figures,…in the paper.)
[当你已经囊括了所有的数据（或者你明确知道你还需要收集哪些额外的数据），有了一个合理的构架，你对这些都感到满意时，将大纲交给我。简要地标明哪些地方还缺数据，你认为（或推测）这些数据大概是什么样。如果你的推测是正确的，你将如何去解释它。拿到你的大纲后，我将把我的观点，建议反馈给你。一般，我们需要四或五个来回才能达成一致（中间经常还需要补做一些实验）。在我们的意见一致后，所有的数据通常以最终（或接近最终的）形式确定下来（也就是说，在提纲中的表格，图表等最终将成为文章中的表格，图表）。] You can then start writing, with some assurance that much of your prose will beused.
The key to efficient use of your and my time is that we start exchangingoutlines and proposals as early in a project as possible. Do not, under anycircumstances, wait until the collection of data is “complete” before startingto write an outline. No project is ever complete, and it saves enormous effortand much time to propose a plausible paper and outline as soon as you see thebasic structure of a project. Even if we decide to do significant additionalwork before seriously organizing a paper, the effort of writing an outline willhave helped to guide the research.
2.3 The outline
What should an outline contain?
3.Abstract: Do not write an abstract. That can be done when the paper iscomplete.
The first paragraph or two should be written out completely. Pay particularattention to the opening sentence. Ideally, it should state concisely theobjective of the work, and indicate why this objective is important.
In general, the Introduction should have these elements:
*The objectives of the work.
*The justification for these objectives: Why is the work important?
*Background: Who else has done what? How? What have we done previously?
*Guidance to the reader. What should the reader watch for in the paper? Whatare the interesting high points? What strategy did we use?
*Summary conclusion. What should the reader expect as conclusion? In advancedversions of the outline, you should also include all the sections that will goin the Experimental section (at this
point, just as paragraph subheadings).
5.Results and Discussion.
The results and discussion are usually combined. This section should beorganized according to major topics. The separate parts should have subheadingsin boldface to make this organization clear, and to help the reader scanthrough the final text to find the parts that interest him or her. Thefollowing list includes examples of the phrases that might plausibly serve assection headings:
*Synthesis of Alkane Thiols
*Characterization of Monolayers
*Absolute Configuration of the Vicinal Diol Unit
*Hysteresis Correlates with Roughness of the Surface
*Dependence of the Rate Constant on Temperature
*The Rate of Self-Exchange Decreases with the Polarity of the Solvent
Try to make these section headings as specific and information-rich aspossible. For example, the phrase “The Rate of Self-Exchange Decreases with ThePolarity of The Solvent” is obviously longer than “Measurement of Rates,” butmuch more useful to the reader. In general, try to cover the major commonpoints:
*Synthesis of starting materials
*Characterization of products
*Methods of characterization
*Methods of measurement
*Results (rate constants, contact angles, whatever)
In the outline, do not write any significant amount of text, but get all thedata in their proper place: any text should simply indicate what will go inthat section.
*Figures (with captions)
*Schemes (with captions and footnotes)
*Tables (correctly formatted)
Remember to think of a paper as a collection of experimental results,summarized as clearly and economically as possible in figures, tables,equations, and schemes. The text in the paper serves just to explain the data,and is secondary. The more information that can be compressed into tables,equations, etc., the shorter and more readable the paper will be.
In the outline, summarize the conclusions of the paper as a list of shortphrases or sentences. Do not repeat what is in the Results section, unlessspecial emphasis is needed. The Conclusions section should be just that, andnot a summary. It should add a new, higher level of analysis, and
shouldindicate explicitly the significance of the work.
Include, in the correct order to correspond to the order in the Resultssection, all of the paragraph subheadings of the Experimental section.
2.4 In summary:[总结：]
* Start writing possible outlines for papers early in a project. Do not waituntil the “end”. The end may never come.
* Organize the outline and the paper around easily assimilated data - tables,equations, figures, schemes - rather than around text.
[整理提纲和论文要围绕易于接受的数据—表格，方程式，图表，示意图，而不是围绕正文。] * Organize in order of importance, not in chronological order. An importantdetail in writing paper concerns the weight to be given to topics. Neophytesoften organize a paper in terms of chronology: that is, they recount theirexperimental program, starting with their cherished initial failures andleading up to a climactic successful finale. This approach is completely wrong.Start with the most important results, and put the secondary results later, ifat all. The reader usually does not care how you arrived at your big results,only what they are. Shorter papers are easier to read than longer ones.
3. Some Points of English Style
1) Do not use nouns as adjectives:
2) The word “this” must always be followed by a noun, so that its reference isexplicit
3) Describe experimental results uniformly in the past tense.
4) Use the active voice whenever possible.
5) Complete all comparisons.
6) Type all papers double-spaced (not single-or one-and-a-half spaced), andleave 1 space after colons, commas, and after periods at the end of sentences.Leave generous margins. (generally,
1.25” on both sides & top &bottom).
Assume that we will write all papers using the style of the American ChemicalSociety. You can get a good idea of this style from three sources:
1) The Journal. Simply look at articles in the journals and copy theorganization you see there.
2) Previous papers from the group. By looking at previous papers, you can seeexactly how a paper should “look”. If what you wrote looks different, itprobably is not what we want.
3) The ACS Style Guide: A Manual for Authors and Editors. (Janet S. Dodd,Editor Washington,
D.C. USA 1997) . Useful detail, especially the section on references.
I also suggest you read Strunk and White, The Elements of Style (Longman: NewYork, 2000, 4th
edition) to get a sense for English usage. Two excellent bookson the design of graphs and figures are: “The Visual Display of QuantitativeInformation” by Edward R. Tufte, Graphics Press (1983)— and “EnvisioningInformation” also by Edward R. Tufte, Graphics Press (1990).